714 
4.3 Errors in Albedo for Various Formulations of and Approximations to Albedo 
Figure 5 Errors in albedo for bare soil (June) 
at 826 nm for atmosphere 1 
Figure 6 Errors in albedo for alkali flat at 826 
nm for atmosphere 1 
Figure 7 Errors in albedo for prairie grass 
(June) at 826 nm for atmosphere 1 
Figure 8 Errors in albedo for desert scrub at 
826 nm for atmosphere 1 
Figures 5 to 8 show the percentage error in ’albedo’ for the five formulations p, to p 5 (marked as numbers 1 
to 5 in the figures) when compared to the albedo calculated from equation (6) (There is no significant difference 
between the results for atmospheres 1 and 2). We can see that the results are grouped into distinct two trends: 
one for the formulations which assume albedo constant with sun angle (p 2 and p 5 ); and one for the rest. It is 
also apparent that the variation between any of the formulations is small, except at very low solar elevations, and 
that all of the models which define various approximations to the ’clear sky’ (i.e. no cloud) albedo provide 
extremely good estimates, except, at low sun angles. Since p 2 is equivalent to the albedo under totally overcast 
skies, we can also conclude that the marked difference (around 15% for the prairie grass, and 5% for the rest at 
90° solar elevation) between this measure and the clear sky albedo requires us to provide estimates of both 
measures to the communities requiring albedo parameters. 
5 - CONCLUSIONS 
• Clear sky albedo can vary significantly as a function of sun angle for a variety of cover types. It is 
not, therefore, generally desirable to model it as a constant unless proportionate errors of up to 15% 
(vegetation) and 5% (desert scrub, alkali flat, bare soil) can be tolerated for this factor alone. • 
• Albedo values calculated for pure diffuse illumination can vary significantly from clear sky values.